I agree with Becker that the settlement between GM and the UAW was a major victory for GM; I suspect that the strike was a face-saving gesture for the union's leadership--a pretense that the strike, that symbol of union power, had wrested important concessions from the company. What is significant about the health trust fund is not that it will remove a large contingent liability from GM's books--a matter simply of accounting--but that it will enable GM to cap that liability by paying a lump sum to create the trust, which then becomes responsible for the employees' (including retired employees') health costs. The risk of soaring health costs is shifted from the company to the workers. If the trust grows faster, through investment of its assets, than the health costs of the trust‚Äôs beneficiaries, the workers will do fine. If not, they, rather than GM, will be the losers. So a cloud of uncertainty over the company's future will be lifted. The purpose of unions is simple: it is to increase unionized workers' incomes above the competitive level, where "income" includes not only wages and fringe benefits but also safety and other aspects of working conditions. The unions do this by cartelizing the sale of labor services. A union that organizes the work force of the major competitors in a market obtains monopoly power over the supply of labor to those firms, and that power enables it to negotiate better terms for its members than they could obtain in a competitive labor market. (The union cannot force the workers in the collective bargaining unit to join the union, but it can charge them an "agency fee" for negotiating on their behalf as well as on behalf of the union members.) The union takes a share of the monopoly gain in the form of union dues. As with any monopoly, union cartelization causes inefficient substitution, whether of capital for labor or of production by nonunionized competitors, and it also generates rent-seeking costs--the costs to unions of organizing an employer's work force and the costs to employers of fighting union organizing efforts. Not only are union cartels inefficient, but they penalize the least productive workers, workers not worth the union wage to employers. What has wrecked the unions, outside of the public employment sector, is the rise of competition in the product markets that used to provide the most favorable conditions for unionization. When thousands of workers are doing essentially the same work in the same plant in an oligopolistic industry, the cost of organizing is minimized. The existence of a large homogeneous work force reduces the cost of communicating the union‚Äôs message and reduces conflicts of interest among workers that would make it difficult to agree to a common package of wages, benefits, and improved working conditions. With few producers, the cost of organizing the entire industry's work force is reduced and the benefits to the workers and to the union (in union dues and agency fees) increased. These conditions are most likely to be found in traditional assembly-line manufacturing, which as Becker notes has declined, in significant part because of the pressure of foreign competition. (That is why unions oppose free trade.) The more competitive an industry, the more difficult it is for unions to extract significant concessions from an employer: higher labor costs will simply deflect an employer's customers to his competitors. Unionization accelerates its own decline. But the particular pickle that GM and the other American auto manufacturers, together with the UMW, find themselves in is also due partly to the U.S. tax code, which makes health benefits deductible to the employer and nontaxable to the employee. This unsound tax policy creates an incentive for the employer and the union to negotiate generous health benefits in lieu of generous wage increases--and without careful controls the employer‚Äôs obligations can skyrocket. "Lifetime" benefits can be a highly attractive perk if health costs rise only gradually, but if they rise rapidly (in part because of increasing longevity, in part because of better but more costly therapies) the employer may find himself saddled with labor costs that make his product noncompetitive. So GM and the other automakers took a big risk in agreeing to extremely generous health benefits, but so did the UMW. For it is difficult for a union to agree to retract concessions that it has obtained from the employers with whom it bargains; it signals weakness. But by hesitating to backpedal, the union has jeopardized the automakers' survival and by doing so has jeopardized its own survival--and guaranteed its decline. Another factor in the decline of unions may, paradoxically, be the National Labor Relations Act, an important New Deal measure that has long been thought pro-union. Before the Act was passed, there was plenty of union activity (the percentage of the work force that was unionized was larger before the Act was passed than it is today), but because of the absence of a tight legal framework for union organizing, labor-management struggles were more like war than like political campaigning. There was a good deal of violence and the organizers, as underdogs, earned a large measure of public sympathy. The NLRA created a detailed legal regime of government-supervised electoral competition and negotiation; violence ceased to be a significant factor in labor relations--and the labor heroes disappeared, and with them unionization as a burning political issue. As Becker notes, the area in which unions have made gains in recent decades is that of public employment. Public employment would always have been a fertile field for union organizing had it not been for laws forbidding strikes by public employees. Public employers provide services rather than products, and service interruptions due to strikes impose greater costs than interruptions in the production of goods because service cannot be produced for inventory or stored; public employees are voters and so government is reluctant to take tough measures against strikers; and public employers are generally monopolists. On all these counts, unionization today poses greater risks to efficiency in the public than in the private sector. In the private sector, the decline in unionization has greatly reduced the power of unions to extract supracompetitive wages and benefits--to the point where, the continued exemption of most union activities from the antitrust laws probably makes good sense. The benefits of antitrust enforcement against practices that nowadays have probably only a small effect on the efficiency of labor markets might well fall short of the costs in enforcement.